Control of extremely fast competitive consecutive reactions using micromixing. Selective Friedel-Crafts aminoalkylation

Article abstract of DOI:10.1021/ja0527424

Friedel-Crafts reactions of aromatic and heteroaromatic compounds with an N-acyliminium ion pool were studied. The reaction of 1,3,5-trimethylbenzene in a batch reactor gave rise to the selective formation of a monoalkylation product (69%). Presumably, th

Full text of DOI:10.1021/ja0527424

Published on Web 07/30/2005
Control of Extremely Fast Competitive Consecutive Reactions
using Micromixing. Selective Friedel-Crafts Aminoalkylation
Aiichiro Nagaki,† Manabu Togai,† Seiji Suga,† Nobuaki Aoki,‡ Kazuhiro Mae,‡ and
Jun-ichi Yoshida*,†
Contribution from the Department of Synthetic Chemistry and Biological Chemistry and
Department of Chemical Engineering, Graduate School of Engineering, Kyoto UniVersity,
Katsura, Kyoto 615-8510
Received April 27, 2005; E-mail: yoshida@sbchem.kyoto-u.ac.jp
Abstract: Friedel-Crafts reactions of aromatic and heteroaromatic compounds with an N-acyliminium ion
pool were studied. The reaction of 1,3,5-trimethylbenzene in a batch reactor gave rise to the selective
formation of a monoalkylation product (69%). Presumably, the second alkylation is slower than the first
alkylation because of the protonation of the monoalkylation product that decreases its reactivity. The reaction
of 1,3,5-trimethoxybenzene, however, gave rise to the formation of both monoalkylation (37%) and
dialkylation (32%) products. Disguised chemical selectivity due to faster reaction than mixing seems to be
responsible for the lack of selectivity. The use of micromixing was found to be quite effective to solve this
problem to increase the selectivity. The monoalkylation product was obtained in 92% yield together with a
small amount of the dialkylation product (4%). The reaction with various aromatic and heteroaromatic
compounds revealed that the low mono/dialkylation selectivity was observed only for highly reactive
aromatics. In such cases, the use of micromixing was quite effective to improve the selectivity. On the
basis of micromixing, the selective sequential dialkylation using two different N-acyliminium ions was
achieved. CFD simulations using a laminar flow and finite-rate model are consistent with the experimental
observations and clearly indicate the importance of mixing.
Scheme 1
Introduction
Chemical reactions are usually achieved by the mixing of
two components such as substrate/reagent and substrate/catalyst.
Molecules of reaction components come together and collide
with each other upon mixing to accomplish the chemical reaction
to give products. Although synthetic organic chemists have
rather neglected the importance of mixing, the outcomes of
chemical reactions are sometimes very sensitive to the way of
mixing.
Let us focus on a single type of reaction, in which the mixing
may affect the product selectivity. The phenomena that two
consecutive reactions occur competitively are popular, and it is
important to favor the first reaction over the second to synthesize
a desired compound efficiently. Product selectivities of such
competitive consecutive reactions are intrinsically determined
by kinetics. However, if the reaction is faster than mixing, simple
kinetics based on the homogeneity of the solution does not work.
It is known that selectivity of competitive consecutive
reactions represented by Scheme 1 is often determined not only
by the rate of two reactions but also by the way of mixing.1
When k1 . k2 and the stoichiometric ratio of two reaction
components is around unity (A:B ) 1:1), the yield of the
primary product P1 should be high. Experiments often contra-
dicted this chemically based prediction. The measured yield of
P1 is sometimes much smaller than that of P2. The selectivity
is not only determined by classical kinetic principles but also
by the way in which the reaction components are mixed.
Rys propesed the concept of disguised chemical selectivities
and pointed out the importance of the way of mixing to control
Mixing in a solution phase is defined as a phenomenon that
creates homogeneity of all species in the solution. Chemists have
tended to believe that such homogeneity is achieved immediately
after the addition of one component to the other before the
chemical reaction is initiated. This is true if the mixing time is
much shorter than the reaction time. In other words, if the
reaction is slower than the mixing, the reaction proceeds in a
homogeneous solution. However, it is sometimes not the case.
Reactions are sometimes faster than mixing, and therefore, the
reaction proceeds before the homogeneity of the solution has
been established. The use of fast reactions is highly desired in
chemical synthesis from a viewpoint of efficiency. Productivity
of faster reactions is obviously better than that of slower
reactions. In such fast reactions, the way of mixing should
strongly affect the outcome of the reaction.
(1) For example, Bourne, J. R. in Handbook of Batch Process Design; Sharratt,
P. N., Ed.; Blackie Adademic and Professional: Dordrecht, The Nether-
lands, 1997; pp 139-152, and references therein.
† Department of Synthetic Chemistry and Biological Chemistry.
‡ Department of Chemical Engineering.
9
11666
J. AM. CHEM. SOC. 2005, 127, 11666-11675
10.1021/ja0527424 CCC: $30.25 © 2005 American Chemical Society

Competitive Consecutive Reactions using Micromixing
A R T I C L E S
Scheme 2
Figure 1. Schematic representation of mixing of the determined reaction
course for competitive consecutive reactions.
and industrial viewpoints.6 The most widely accepted definition
of microreactors is “reactors having microstructures for chemical
reactions”. According to this definition, microreactors are not
necessary small devices in total size. They can be relatively
large as far as they contain microstructures. It is also important
to note that microreactors do not necessarily mean reactors for
the production of a small quantity of compounds. Microreactors
can be applied for industrial scale production.7
With the advancement of microfabrication technology to
make various kinds of microreactors including micromixers and
microheat exchangers, the development of new synthetic
methodologies based upon the inherent features that exist at
the micrometer scale is strongly needed. It is generally expected
that extremely fast and exothermic reactions can be conducted
in a highly controlled manner in microreactors by virtue of the
advantages of efficient mixing and heat transfer. In this paper,
we focus on the efficient mixing ability of micromixers8 to solve
the problem of disguised selectivity of extremely fast competi-
tive consecutive reactions.9
chemical reactions.2 Rys wrote “If we add a solution of species
B to a solution of species A, eddies of solution B in solution A
are created. As a first approximation, these eddies can be
considered as spherical drops with constant mean radius R. The
lifetime of such an eddy can be estimated to be 0.01-1 s. The
radius R depends on the intensity of the turbulence created by
mixing and may be controlled, for example, by mechanical
stirring. From the theory of turbulence, one can estimate the
minimum mean size of such elements of liquid. For the common
solvents water, methanol, and ethanol, the mean minimum radius
R of the eddies in optimal turbulence is approximately 10-2 to
10-3 cm.”2a Rys also wrote, “After a first time interval, A and
B will have reacted at the periphery of an eddy to produce the
primary product P1. In a further time interval, a molecule B
can react with a further molecule A to form P1, but only if it
succeeds in diffusing through the peripheral zone of P1
molecules already formed without being trapped there by the
substance P1 in a reaction affording the secondary product P2.
This succeeds less often, the longer the relaxation time of the
diffusion process is in comparison to the relaxation time of the
secondary reaction. In the extreme case, a mixing controlled
reaction will convert all the P1 into P2 before the molecule B
finds a further molecule A. Thus, at the end of the reaction
practically only the secondary product P2 is present, and no
primary product P1 can be detected. In this case, the selectivity
kl/k2 loses its influence on the product distribution.”2b
As an example of competitive consecutive reactions, the
Friedel-Crafts reaction is very popular in organic chemistry
(Scheme 2).10 In the case where the R group is an electron-
(5) For organic synthesis (a) Salimi-Moosavi, H.; Tang, T.; Harrison, D. J. J.
Am. Chem. Soc. 1997, 119, 8716. (b) Chambers, R. D.; Spink, R. C. H.
Chem. Commun. 1999, 883. (c) de Bellefon, C.; Tanchoux, N.; Caravieilhes,
S.; Grenouillet, P.; Hessel, V. Angew. Chem., Int. Ed. 2000, 39, 3442. (d)
Suga, S.; Okajima, M.; Fujiwara, K.; Yoshida, J. J. Am. Chem. Soc. 2001,
123, 7941. (e) Yoshida, J.; Suga, S. Chem.sEur. J. 2002, 8, 2651. (f) Watts,
P.; Wiles, C.; Haswell, S. J.; Pombo-Villar, E.; Styring, P. Chem. Commun.
2001, 990. (g) Hisamoto, H.; Saito, T.; Tokeshi, M.; Hibara, A.; Kitamori,
T. Chem. Commun. 2001, 2662. (h) Sands, M.; Haswell, S. J.; Kelly, S.
M.; Skelton, V.; Morgan, D. O.; Styring, P.; Warrington, B. Lab Chip 2001,
1, 64. (i) Wiles, C.; Watts, P.; Haswell, S. J.; Pombo-Villar, E. Lab Chip
2001, 1, 100. (j) Wiles, C.; Watts, P.; Haswell, S. J.; Pombo-Villar, E.
Chem. Commun. 2002, 1034. (k) Fukuyama, T.; Shinmen, M.; Nishitani,
S.; Sato, M.; Ryu, I. Org. Lett. 2002, 4, 1691. (l) Ueno, M.; Hisamoto, H.;
Kitamori, T.; Kobayashi, S. Chem. Commun. 2003, 936. (m) Garcia-Egido,
E.; Spikmans, V.; Wong, S. Y. F.; Warrington, B. H. Lab Chip 2003, 3,
73. (n) Lai, S. M.; Martin-Aranda, R.; Yeung, K. L. Chem. Commun. 2003,
218. (o) Mikami, K.; Yamanaka, M.; Islam, M. N.; Kudo, K.; Seino, N.;
Shinoda, M. Tetrahedron Lett. 2003, 44, 7545. (p) Kobayashi, J.; Mori,
Y.; Okamoto, K.; Akiyama, R.; Ueno, M.; Kitamori, T.; Kobayashi, S.
Science 2004, 304, 1305. (q) Wu, T.; Mei, Y.; Cabral, J. T.; Xu, C.; Beers,
K. L. J. Am. Chem. Soc. 2004, 126, 9980. (r) Nagaki, A.; Kawamura, K.;
Suga, S.; Ando, T.; Sawamoto, M.; Yoshida, J. J. Am. Chem. Soc. 2004,
126, 14702. (s) Horcajada, R.; Okajima, M.; Suga, S.; Yoshida, J. Chem.
Commun. 2005, 1303.
(6) Reviews: for example, (a) DeWitt, S. H. Curr. Opin. Chem. Biol. 1999, 3,
350. (b) Okamoto, H. J. Synth. Org. Chem., Jpn. 1999, 57, 805. (c)
Sugawara, T. Pharmacia 2000, 36, 34. (d) Wo¨rz, O.; Ja¨ckel, K. P.; Richter,
Th.; Wolf, A. Chem. Eng. Sci. 2001, 56, 1029. (e) Haswell, S. J.; Middleton,
R. J.; O’Sullivan, B.; Skelton, V.; Watts, P.; Styring, P. Chem. Commun.
2001, 391. (f) de Mello, A.; Wooton, R. Lab Chip 2002, 2, 7N. (g) Fletcher,
P. D. I.; Haswell, S. J.; Pombo-Villar, E.; Warrington, B. H.; Watts, P.;
Wong, S. Y. F.; Zhang, X. Tetrahedron 2002, 58, 4735. (h) Ja¨hnisch, K.;
Hessel, V.; Lowe, H.; Baerns, M. Angew. Chem., Int. Ed. Engl. 2004, 43,
406.
As described previously, the kinetically based selectivity may
be disguised by the mixing problem. This phenomenon may
occur often in macroscale batch reactors, where the mixing is
not so fast. As a matter of fact, several examples have already
been reported in the literature.3 We envisioned that the problem
of disguised chemical selectivity can be solved by extremely
fast mixing using a microfabricated device, a micromixer.
Microfabricated devices for chemical reactions are generally
called microreactors,4 and they are expected to make a
revolutionary change in chemical synthesis5 from both academic
(2) (a) Rys, P. Acc. Chem. Res. 1976, 10, 345. (b) Rys, P. Angew. Chem., Int.
Ed. Engl. 1977, 12, 807. In the original papers of Rys A is added to B, but
here we consider the reverse addition (addition of B to A).
(3) (a) Agre, C. L.; Dinga, G.; Pflaum, R. J. Org. Chem. 1956, 21, 561. (b)
Stoutland, O.; Helgen, L.; Agre, C. L. J. Org. Chem. 1959, 24, 818. (c)
Tolgyesi, W. S. Can. J. Chem. 1965, 43, 343. (d) Wood, P. B.; Higginson,
W. C. E. J. Chem. Soc. A 1966, 1645. (e) Hanna, S. B.; Hunziker, E.;
Saito, T.; Zollinger, H. HelV. Chim. Acta 1969, 52, l537. (f) Paul, E. L.;
Treyba1, R. E. Am. IChemE J. 1971, 17, 718. (g) Ridd, J. H. Acc. Chem.
Res. 1971, 4, 248. (h) Hunziker, E.; Penton, J. R.; Zollinger, H. HelV. Chim.
Acta 1971, 54, 2043. (i) Bourne, J. R.; Rys, P.; Suter, K. Chem. Eng. Sci.
1977, 32, 711. (j) Bourne, J. R.; Kozicki, F.; Chem. Eng. Sci. 1977, 32,
1538. (k) Kaminski, R.; Lauk, U.; Skraba1, P.; Zollinger, H. HelV. Chim.
Acta 1983, 66, 2002. (l) Jacobson, A. R.; Makris, A. N.; Sayre, L. M. J.
Org. Chem. 1987, 52, 2592. (m) Paul, E. L.; Mahadevan, H.; Foster, J.;
Kennedy, M.; Midler, M. Chem. Eng. Sci. 1992, 47, 2837.
(4) (a) Ehrfeld, W. Microreaction Technology; Springer: Berlin, 1998. (b)
Ehrfeld, W.; Hessel, V.; Lo¨we, H. Microreactors; Wiley-VCH: Weinheim,
2000. (c) Manz, A.; Becker, H. Microsystem Technology in Chemistry and
Life Sciences; Springer: Berlin, 1999. (d) Jensen, K. F. Chem. Eng. Sci.
2001, 56, 293. (e) Hessel, V.; Hardt, S.; Lo¨we, H. Chemical Micro Process
Engineering; Wiley-VCH: Weinheim, 2004.
(7) The use of microreactors toward industrial production: (a) Iwasaki, T.;
Yoshida, J. Macromolecules 2005, 38, 1159. (b) Kawaguchi, T.; Miyata,
H.; Ataka, K.; Mae, K.; Yoshida, J. Angew. Chem., Int. Ed. 2005, 44, 2413.
(8) For example, (a) Kakuta, M.; Bessoth, F. G.; Manz, A. Chem. Record 2001,
1, 395. (b) Stroock, A. D.; Dertinger, S. K. W.; Ajdari, A.; Meziæ, I.;
Stone, H. A.; Whitesides, G. M. Science 2002, 295, 647. (c) Johnson, T.
J.; Ross, D.; Locascio, L. E. Anal. Chem. 2002, 74, 45. (d) Seong, G. H.;
Crooks, R. M. J. Am. Chem. Soc. 2002, 124, 13360. (e) Song, H.; Tice, J.
D.; Ismagilov, R. F. Angew. Chem., Int. Ed. 2003, 42, 768. (f) Zheng, B.;
Roach, L. S.; Ismagilov, R. F. J. Am. Chem. Soc. 2003, 125, 11170. (g)
Song, H.; Ismagilov, R. F. J. Am. Chem. Soc. 2003, 125, 14613.
9
J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005 11667

A R T I C L E S
Nagaki et al.
Scheme 3
donating group such as an alkyl group (Friedel-Crafts alkyla-
tion), it is well-recognized that the primary product is, in
principle, more reactive than the starting aromatic compound.11
Thus, the second reaction takes place more easily than the first
reaction. This inevitable problem is usually avoided by the use
of a large excess amount of a starting aromatic compound
(sometimes an aromatic compound is used as a solvent).
Selective monoalkylation using 1 equiv of an aromatic com-
pound has remained one of the most challenging problems in
organic synthesis for many years. If the R group is an electron-
withdrawing group such as an acyl group (Friedel-Crafts
acylation), the primary product is less reactive than the parent
aromatic compound. In this case, the second reaction is slower
than the first reaction, and therefore, it is possible to stop the
reaction to obtain the primary product selectively. Even if the
R group is an electron-withdrawing group, the second reaction
may sometimes take place in a significant extent because of
the problem of disguised chemical selectivity owing to slow
mixing.3 In this case, extremely fast micromixing has a chance
to improve product selectivity.
Scheme 4
Scheme 5
In the Friedel-Crafts reactions, carbocations12 are used as
reagents, which are usually generated in situ by Lewis acid
promoted reactions of their precursors such as alkyl halides and
acyl halides. In such cases, however, the generation process is
reversible, and the equilibrium usually lies to the precursor.
Recently, we have developed the cation pool method that
involves irreversible oxidative generation and accumulation of
carbocations in the absence of nucleophiles at low temperature.5e,13
For example, N-acyliminium ion pools can be easily generated
by anodic oxidation14 of either corresponding carbamates or
carbamates having a silyl group as an electroauxiliary15 and
accumulated in a solution at low temperature (Scheme 3). We
envisioned that the use of N-acyliminium ion pools in Friedel-
Crafts reactions would serve as a good probe to test the problem
of disguised chemical selectivity and the effectiveness of
extremely fast micromixing.
Thus, we have examined the Friedel-Crafts reactions of
aromatic compounds with N-acyliminium ion pools and found
that disguised chemical selectivity was observed in some cases
and that this problem was solved by the micromixing.16
Results and Discussion
Generation of N-Acyliminium Ion Pools. We chose to study
N-acyliminium ion 2 that does not have a substituent at the
iminium carbon as shown in Scheme 4 because less sterically
demanding N-acyliminium ions seem to be suitable for the study
of monoalkylation/polyalkylation selectivity. N-Acyliminium ion
2 was generated by the anodic oxidation of carbamate 1 having
a silyl group as an electroauxiliary in Bu4NBF4/CH2Cl2 and
accumulated in a solution in the absence of a nucleophile. The
formation of 2 as a single species was indicated by NMR (1H
NMR: 8.56 and 8.83 ppm due the methylene protons, 13C
NMR: 177.0 ppm due to the methylene carbon).
(9) It is reported that the product distribution of two competitive reactions is
influenced by the effectiveness of micromixing: (a) Ehrfeld, W.; Golbig,
K.; Hessel, V.; Lo¨we, H.; Richter, T. Ind. Eng. Chem. Res. 1999, 38, 1075.
(b) Suga, S.; Nagaki, A.; Tsutsui, Y.; Yoshida, J. Org. Lett. 2003, 5, 945.
(c) Yoshida, J.; Nagaki, A.; Iwasaki, T.; Suga, S. Chem. Eng. Technol.
2005, 28, 259. (d) Hessel, V.; Hofmann, C.; Lo¨we, H.; Meudt, A.; Scherer,
S.; Scho¨nfeld, F.; Werner, B. Org. Process Res. DeV. 2004, 8, 511.
(10) (a) Olah, G. A.; Spear, R. J.; Messina, G.; Westerman, P. W. J. Am. Chem.
Soc. 1975, 97, 4051. (b) Olah, G. A. J. Am. Chem. Soc. 1965, 87, 1103.
(c) Olah, G. A.; Schlosberg, R. H.; Porter, R. D.; Mo, Y. K.; Kelly, D. P.;
Mateescu, G. D. J. Am. Chem. Soc. 1972, 94, 2034. (d) Effenberger, F.
Acc. Chem. Res. 1989, 22, 27. (e) Olah, G. A. Acc. Chem. Res. 1971, 4,
2040.
N-Acyliminium ion 4 was also generated from precursor 3
in a similar fashion as shown in Scheme 5.
Reactivity of N-Acyliminium Ion Pools (2) and (4) with
Allyltrimethylsilane. Although we have already developed the
chemistry of N-acyliminium ions having a substituent on the
iminium carbon, the reactivity of N-acyliminium ions that do
not have a substituent at the iminium carbon is not well-known.
(11) For example, Bruckner, R. AdVanced Organic Chemistry: Reaction
Mechanism; Harcourt/Academic Press: San Diego, 2002.
(12) For example, (a) Prakash, G. K. S.; Schleyer, P. v. R. Stable Carbocation
Chemistry; Wiley: New York, 1997. (b) Olah, G. A.; Laali, K. K.; Wang,
Q.; Prakash, G. K. S. Onium Ions; Wiley: New York, 1998.
(13) (a) Yoshida, J.; Suga, S.; Suzuki, S.; Kinomura, N.; Yamamoto, A.;
Fujiwara, K. J. Am. Chem. Soc. 1999, 121, 9546. (b) Suga, S.; Suzuki, S.;
Yamamoto, A.; Yoshida, J. J. Am. Chem. Soc. 2000, 122, 10244. (c) Suga,
S.; Okajima, M.; Yoshida, J. Tetrahedron Lett. 2001, 42, 2173. (d) Suga,
S.; Okajima, M.; Fujiwara, K.; Yoshida, J. J. Am. Chem. Soc. 2001, 123,
7941. (e) Suga, S.; Suzuki, S.; Yoshida, J. J. Am. Chem. Soc. 2002, 124,
30. (f) Yoshida, J.; Suga, S. Chem.sEur. J. 2002, 8, 2650. (g) Suga, S.;
Watanabe, M.; Yoshida, J. J. Am. Chem. Soc. 2002, 124, 14824. (h) Suga,
S.; Kageyama, Y.; Babu, G.; Itami, K.; Yoshida, J. Org. Lett. 2004, 6,
2709. (i) Szuki, S.; Matsumoto, K.; Kawamura, K.; Suga, S.; Yoshida, J.
Org. Lett. 2004, 6, 3755. (j) Suga, S.; Suzuki, S.; Maruyama, T.; Yoshida,
J. Bull. Chem. Soc. Jpn. 2004, 77, 1545. (k) Suga, S.; Nishida, T.; Yamada,
D.; Nagaki, A.; Yoshida, J. J. Am. Chem. Soc. 2004, 126, 14338.
(14) (a) Shono, T.; Matsumura, Y.; Tsubata, K. J. Am. Chem. Soc. 1981, 103,
1172. (b) Shono, T.; Matsumura, Y.; Tsubata, K. Tetrahedron Lett. 1981,
22, 2411. (c) Moeller, K. D. Tetrahedron 2000, 56, 9527. (d) Kim, S.;
Hayashi, K.; Kitano, Y.; Tada, M.; Chiba, K. Org. Lett. 2002, 4, 3735. (e)
Wanyoike, G. N.; Onomura, O.; Maki, T.; Matsumura, Y. Org. Lett. 2002,
4, 1875.
(15) (a) Yoshida, J.; Nishiwaki, K. J. Chem. Soc., Dalton Trans. 1998, 2589.
(b) Yoshida, J.; Sugawara, M.; Tatsumi, M.; Kise, N. J. Org. Chem. 1998,
63, 5950. (c) Yoshida, J.; Watanabe, M.; Toshioka, H.; Imagawa, M.; Suga,
S. Chem. Lett. 1998, 1011. (d) Sugawara, M.; Mori, K.; Yoshida, J.
Electrochim. Acta 1997, 42, 1995. (e) Yoshida, J.; Ishichi, Y.; Isoe, S. J.
Am. Chem. Soc. 1992, 114, 7594. (f) Yoshida, J.; Maekawa, T.; Murata,
T.; Matsunaga, S.; Isoe, S. J. Am. Chem. Soc. 1990, 112, 1962. (g) Yoshida,
J.; Isoe, S. Tetrahedron Lett. 1987, 28, 6621. (h) Sun, H.; Moeller, K. D.
Org. Lett. 2002, 4, 1547. (i) Kamada, T.; Oku, A. J. Chem. Soc., Perkin
Trans. 1 2002, 1105. Photochemical reactions: (j) Xu, W.; Zhang, X. M.;
Mariano, P. S. J. Am. Chem. Soc. 1991, 113, 8863. (k) Gutenberger, G.;
Steckhan, E.; Blechert, S. Angew. Chem., Int. Ed. 1998, 37, 660. Chemical
reactions: (l) Narasaka, K. Pure Appl. Chem. 1997, 69, 601. (m) Pandey,
G.; Sahoo, A. K.; Gadre, S. R.; Bagul, T. D.; Phalgune, U. D. J. Org.
Chem. 1999, 64, 4990. (n) Sun, H.; Moeller, K. D. Org. Lett. 2003, 5,
3189.
(16) Preliminary results have been reported: Suga, S.; Nagaki, A.; Yoshida, J.
Chem. Commun. 2003, 354.
9
11668 J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005

Competitive Consecutive Reactions using Micromixing
A R T I C L E S
Scheme 6
Scheme 7
Figure 2. Energy diagram of the reaction of the N-methoxycarbonyl-N-
methyliminium ion with 1,3,5-trimethylbenzene.
Scheme 8
however, contradicted with the expectation. The selective
monoalkylation can be explained as follows: Friedel-Crafts
reaction of 8 with 2 generates a proton, which protonates the
monoalylation product 9. The protonated carbamate group
should be strongly electron-withdrawing. Therefore, the proto-
nated form of 9 is considered to be less reactive than 8. Thus,
the second alkylation is slower than the first alkylation. This is
consistent with the selective formation of the monoalkylation
product 9.
NMR Studies. To get the evidence for the protonation of 9,
13C NMR studies were carried out. The nuetral form of 9
exhibited the signal due to the carbonyl carbon at 155.0 and
155.4 ppm (two rotamers). The addition of HBF4OMe2 caused
a downfield shift of the signal (157.5 and 157.7 ppm). This is
consistent with the protonation at the carbonyl oxygen, which
made the carbonyl carbon more electron-deficient. The 13C
NMR spectrum of the product solution obtained by the reaction
of 2 and 8 exhibited the signal at 157.6 and 157.8 ppm. These
observations strongly suggest that the protonated form of 9 was
produced by the reaction of 2 and 8.
DFT Calculations. To obtain deeper insight into the mech-
anism of the present reaction, DFT calculations (B3LYP/6-31G-
(d)) were carried out for a model system (N-methoxycarbonyl-
N-methyliminium ion with 1,3,5-trimethylbenzene) shown in
Figure 2 in the gas phase. The calculations indicate that complex
A, which is formed prior to the reaction, is converted into the
Wheland complex C through transition state B. In the next step,
complex C is converted to E through transition state D, where
the proton on the benzene ring is transferred to the carbonyl
oxygen atom. Therefore, E is the protonated form of the final
product. This is consistent with the speculation described
previously. The present reaction is exothermic (ca. 14 kcal/mol),
and the activation energy is very small (ca. 0.3 kcal/mol). These
data obtained by the calculations indicate that the present
reaction is very fast.
Thus, we first examined the reaction of 2 with simple carbon
nucleophiles. The reaction of 2 with allyltrimethylsilane took
place spontaneously at -78 °C to give the corresponding
allylated product (5) in 72% yield as shown in Scheme 6,
indicating that 2 is a highly reactive species. N-Acyliminium
ion 2 also reacted with 3-trimethylsilylcyclohexene to give the
corresponding C-C bond formation product (6) in 72% yield
(Scheme 6). We also examined the reaction of N-acyliminium
ion pool (4) with allyltrimethylsilane. The corresponding
allylated product (7) was obtained in 57% yield, indicating that
4 is also a highly reactive species (Scheme 7).
N-Acyliminium Ion Pool (2) with 1,3,5-Trimethylbenzene
(8) using a Macroscale Batch Reactor. In the study of the
monoalkylation/polyalkylation selectivity of Friedel-Crafts
reaction, we first examined the reaction of 1,3,5-trimethylben-
zene (8) as an aromatic compound because it is highly reactive
and has three separate positions for the attack of an electrophile.
Because the yield of 2 from 1 is estimated as ca. 80% based on
the reactions with various nucleophiles, hereafter 1.2 equiv of
1 was used when 1.0 equiv of 2 was needed for the Friedel-
Crafts reaction. The yields of the products are based on the
amount of the aromatic compounds used.
The reaction in a conventional macroscale batch reactor with
magnetic stirring, which is usually used for laboratory organic
synthesis, was examined. The addition of 2 (generated from
1.2 equiv of 1) to a stirred solution of 8 in dichloromethane
resulted in the selective formation of monoalkylation product
9 (69% yield) (Scheme 8). The dialkylation product 10 was
not detected. It is well-known that the monoalkylation product
is more reactive than the parent compound because alkyl groups
are electron-donating. Therefore, it is generally difficult to stop
the reaction at the monoalkylation stage. The experimental result,
Reaction of N-Acyliminium Ion Pool (2) with 1,3,5-
Trimethoxybenzene (11) using a Batch Reactor. The results
obtained with N-acyliminium ion (2) and trimethylbenzene (8)
indicate that the first alkylation is faster than the second
alkylation and that the disguised chemical selectivity was not
observed. Thus, we examined the reactions of 2 with more
reactive aromatic compounds.
9
J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005 11669

A R T I C L E S
Scheme 9
Nagaki et al.
Since alkoxyl groups are more electron-donating than alkyl
groups, we chose to study 1,3,5-trimethoxybenzene (11) as an
aromatic compound. The addition of 2 (generated from 1.2 equiv
of 1) to a stirred solution of 11 in dichloromethane in a
macroscale batch reactor resulted in the formation of an
essentially 1:1 mixture of monoalkylation product 1217 and
dialkylation product 13 (Scheme 9) after quenching by the
triethylamine.18 The reverse addition gave essentially the same
result. The simultaneous addition of a solution of 2 and that of
11 to a batch reactor also gave a similar result. These
observations contrast sharply with the case of 1,3,5-trimethyl-
benzene. Since the monoalkylation product 12 should be
protonated to deactivate the aromatic ring, the second alkylation
should be slower than the first alkylation. Therefore, the
selectivity observed here is considered to be disguised chemical
selectivity, which is ascribed to faster reaction than mixing.
Reaction of N-Acyliminium Ion Pool (2) with 1,3,5-
Trimethoxybenzene (11) using a Micromixer. Thus, we
examined the reaction using a micromixer to solve the problem
of disguised chemical selectivity. There are several principles
of micromixers such as injection of many small substreams of
one component in a main stream of another component, decrease
of diffusion path perpendicular to the flow direction, manifold
splitting and recombination, and periodic injection of small fluid
segments.
Figure 3. Splitting and recombination type micromixer (YM-1): (a)
principle, (b) outside, and (c) schematic diagram of mixing element.
In the present study, we examined three different types of
micromixers. The first one is a T-shaped tube mixer, and its
diameter is 500 µm. The second one, a Yamatake YM-1 mixer
developed by Mae, is a splitting and recombination type
micromixer. A multilamination type mixer, IMM (Institut fu¨r
Mikrotechnik Mainz GmbH) single mixer, was also examined.
In the YM-1 micromixer (Figure 3), two fluids to be mixed
were introduced into a segment and combined. Then, the mixture
was split into two streams, which were introduced into the next
segment.
In the IMM mixer, the fluids to be mixed are introduced into
the mixing element as two counter-flows, and the fluids stream
into an interdigital channel (25 µm) configuration (Figure 4).
In the next stage, a periodical flow configuration consisting of
the lamellae of the two fluids is generated by means of the slit-
Figure 4. Multilamination type micromixer (IMM single mixer, channel
width ) 25 µm): (a) principle, (b) outside, and (c) schematic diagram of
mixing element.
shaped interdigital channel. Then, the lamellated flow leaves
the device perpendicular to the direction of the feed flows.
The reaction of the N-acyliminium ion 2 with aromatic
compound 11 using these micromixers was carried out as
follows. The solutions of 2 (produced from 1.2 equiv of 1) and
11 were introduced to the mixer by a syringe pumping technique
(flow rate: 5.0 mL/min) at -78 °C, and the product solution
that left from the device was immediately quenched with
triethylamine to avoid further reactions.18
When the T-shaped mixer was used, the selectivity was
essentially the same as the batch reactor (Scheme 10). The use
of the YM-1 mixer increased the selectivity, but still a significant
amount of dialkylation product 13 was formed. We found,
however, that the monoalkylation product 12 was obtained in
excellent selectivity when the IMM micromixer was used. The
(17) R-Amidoalkyation: Zaugg, H. E.; Martin, W. B Org. React. 1965, 14, 52.
(18) The separate reaction of 2 with triethylamine gave rather inactive species
that did not give the allylated product upon treatment with allyltrimethyl-
silane, although the details are not clear at present. Triethylamine might
also suppress the acid-promoted decomposition of the product.
9
11670 J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005

Competitive Consecutive Reactions using Micromixing
A R T I C L E S
Scheme 10
Table 3. Reaction of 2 and 11 in Various Molar Ratios using IMM
Micromixera
1 (precursor of 2, equiv)
12 (%)
13 (%)
1.0
1.2
1.5
2.0
2.5
66
92
92
71
62
2
4
5
19
30
a A solution of 11 (0.417 mmol) in CH2Cl2 (10 mL, -78 °C) and a
cation pool (2, 10 mL, -78 °C) generated from 1 (0.500 mmol) was
simultaneously introduced to the micromixer using syringe pumps.
IMM reported that the mixing efficiency strongly depends on
the flow rate and decreases with a decrease in the flow rate.9a
We also examined the effect of the ratio of the two reaction
components (2 and 11) on the selectivity using the IMM
micromixer. As shown in Table 3, the use of 1.2 equiv of
precursor 1 gave the best results. The use of 2.0 and 2.5 equiv
of 1 resulted in the formation of a significant amount of 13.
Surplus 2 further reacted with 12 to give 13.
It should be noted that the use of 2.5 equiv of 1 (more than
2 equiv of 2) did not give the dialkylation product 13 selectively.
This observation seems to indicate that the reaction of monoalkyl-
ation product 12 (protonated 12) with N-acyliminium ion 2 is
slower than the reaction of the parent aromatic compound 11
with 2.
Table 1. Temperature Effect of the Reaction of 2 with 11
temperature (
°
C)
reactor
conversion of 11 (%)
12 (%)
13 (%)
-78
micromixera
batch reactorb
micromixera
batch reactorb
micromixera
batch reactorb
micromixera
batch reactorb
91
75
92
86
85
99
93
92
37
84
7
70
1
4
32
15
22
19
7
-47
-27
0
30
0
15
1
quantitative
Reaction of N-Acyliminium Ion Pool (2) with Various
Aromatic and Heteroaromatic Compounds. The Friedel-
Crafts reactions of other aromatic compounds with 2 were
examined, and the results are summarized in Table 4.
For less reactive aromatic compounds, such as toluene,
p-xylene, and 1,3,5-trimethylbenezene, the monoalkylation
products were selectively obtained even in a macroscale batch
reactor. The protonation at the carbonyl oxygen, which deceler-
ates the second alkylation as described in the previous section,
seems to be responsible.
The dramatic effect of micromixing is generally observed for
the alkylation of highly reactive aromatic compounds, such as
methoxybenezene, 1,2-dimethoxybenzene, and 1,3,5-trimethoxy-
benzene. The selectivity observed for a macroscale batch reactor
seems to be the disguised chemical selectivity. The use of the
micromixer solved the problem, and the monoalkylation product
was obtained in high selectivity.
The outcome of the reaction with heteroaromatic compounds
was also affected by the way of mixing. The reaction of 2 with
thiophene with the micromixer took place smoothly to give the
monoalkylation product exclusively, while the reaction in the
macroscale batch reactor gave a significant amount of dialkyl-
ation product. A similar tendency was also observed for the
reaction of furan and N-methylpyrrole. These results indicate
that the reactions of 2 with these heteroaromatic compounds
also suffer from the problem of the disguised chemical selectiv-
ity in the macroscale batch reactor and that such problem can
be solved by micromixing.
a IMM single mixer was used. b A 50 mL flask was used.
Table 2. Effect of Flow Rate of the Reaction of 2 with 11 using
the IMM Micromixer
flow rate (mL/min)a
12 (%)
13 (%)
1
3
5
14
52
92
19
14
4
a A solution of 11 (0.417 mmol) in CH2Cl2 (10 mL, -78 °C) and a
cation pool (2, 10 mL, -78 °C) generated from 1 (0.500 mmol) was
simultaneously introduced to the micromixer using syringe pumps.
amount of dialkylation product 13 was very small (12:13 )
96:4, total yield 96%).
The reaction temperature was also found to be an important
factor governing the selectivity (Table 1). In the case of the
IMM micromixer, the selectivity decreased dramatically with
the increase of the temperature. A similar tendency was also
observed for a batch reactor. Another important point is that
the total yield of 12 and 13 also decreased with the increase of
the temperature, although the conversion remained very high.
Such an effect is much larger for the macroscale batch reactor.
Presumably, the products decomposed under the reaction
conditions at higher temperatures. In the case of the micromixer,
better temperature control was achieved by virtue of high
surface-to-volume ratio of the microstructure. In the case of the
batch reactor, however, the temperature control was less
effective, and therefore, significant amounts of the products
decomposed at higher temperatures.
It should be noted here that highly nucleophilic aromatic and
heteroaromatic compounds suffer from the disguised chemical
selectivity. The nucleophilicity parameters reported by Mayr19
Flow rate is also a very important factor that governs the
selectivity of the present reaction. With flow rates of 5 mL/
min, the best result was obtained among those examined.
Decreasing the flow rate caused a decrease of the yield and
selectivity as shown Table 2. The mixing efficiency seems to
be strongly dependent on the flow rate. As a matter of fact,
(19) (a) Mayr, H.; Patz, M. Angew. Chem., Int. Ed. 1994, 33, 938. (b) Gotta,
M. F.; Mayr, H. J. Org. Chem. 1998, 63, 9769. (c) Mayr, H.; Bug, T.;
Gotta, M. F.; Hering, N.; Irrgang, B.; Janker, B.; Kempf, B.; Loos, R.;
Ofial, A. R.; Remennikov, G.; Schimmel, H. J. Am. Chem. Soc. 2001, 123,
9500. (d) Mayr, H.; Kempf, B.; Ofial, A. R. Acc. Chem. Res. 2003, 36, 66.
9
J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005 11671

A R T I C L E S
Nagaki et al.
Table 4. Reaction of N-Acyliminium Ion Pool (2) with Various
Scheme 11
Aromatic and Heteroaromatic Compoundsa
was examined. The first alkylation was carried out with
N-acyliminium ion 2 using the IMM micromixer to obtain the
monoalkylation product 14, which was directly subjected to the
second alkylation with a different N-acyliminium ion 4 in a batch
reactor to obtain dialkylation product 15 in 64% yield (Scheme
11). Therefore, the present method serves as a simple and
straightforward method for the selective introduction of two
different alkyl groups on an aromatic ring.
Estimation of Relative Rates of the First Alkylation and
Second Alkylation. In the previous sections, we assumed that
the second alkylation was slower than the first alkylation because
the carbamate group in the monoalkylation product should be
protonated to deactivate the aromatic ring. To confirm this
assumption, we carried out the following experiments using a
microflow system shown in Figure 5.
A solution of an aromatic compound and the cation pool 2
(generated from two equivalents of the precursor 1) was mixed
with a YM-1 micromixer at -78 °C (flow rate: 8 mL/min),
and the resulting mixture was introduced to a microtube reactor
(diameter: 1.0 mm and length: 1.0 cm). Triethylamine was
introduced through the second micromixer (YM-1) (flow rate:
8 mL/min) to quench the reaction. The results are summarized
in Table 5.
The reaction of 1,3,5-trimethoxybenzene (11) was completed
within the residence time of 0.03 s. Monoalkylation product 12
was obtained in 78% yield together with dialkylation product
13 (19%). The reaction of monoalkylation product 12, which
was synthesized and isolated in a separate experiment, with 2
a The reactions were usually carried out with 2, which was generated
from 0.50 mmol of 1 and 0.42 mmol of an aromatic compound. b Nucleo-
philicity parameters reported by Mayr.19 c A mixture of two- and four-
substituted products. d The yield was determined by 1H NMR using an
internal standard. e A mixture of two- and three-substituted products (76:
24). f A mixture of two- and three-substituted products (72:28).
Figure 5. Microflow system for the estimation of relative rates of the
reaction of aromatic compounds with the cation pool.
Table 5. Relative Rates of the Reactions of 11 and 12 with 2a
serve as good indicators. Compounds having a nucleophilicity
number (N) higher than ca. -1 suffer from the problem of
disguised selectivity. In such cases, the micromixing serves as
a solution to the problem.
Sequential Dialkyaltion with Two Different N-Acylimin-
ium Ion Pools. To demonstrate the synthetic utility of the
selective Friedel-Crafts monoalkylation using a micromixer,
the sequential alkylation with two different alkylating agents
residence
time (s)
monoalkylation
product 12 (%)
dialkylation
product 13 (%)
aromatic compound
11
0.03
0.03
0.03
0.30
78
9
56
18
19
56
6
12 (isolated)
12 (without isolation, protonated)
12 (without isolation, protonated)
56
a A microsystem consisting of two micromixers (M1 and M2) and a
tube reactor (R1) was used (Figure 5).
9
11672 J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005

Competitive Consecutive Reactions using Micromixing
A R T I C L E S
Scheme 12
Figure 6. Models of micromixer for CFD simulation. Mixer (a): W )
100 µm. Mixer (b): W ) 25 µm. Mixer (c): W ) 2.5 µm. The product
yields and distribution were calculated when the reaction is complete. Results
are summarized in Tables 6-8.
Table 6. Product Yields and Distribution Obtained by CFD
Simulation (k1 ) 104 L/(mol s) and k2 ) 102 L/(mol s))
also proceeded smoothly. The reaction rate seems to be
comparable to that of 11 because most of 12 was consumed
within the residence time of 0.03 s. This observation indicates
that the reactivity of unprotonated monoalkylation product 12
is similar to 11. The reaction of a solution of 12, which was
generated by the reaction of 11 with 2 using micromixing in a
separate experiment, was found to be much slower. A significant
amount of 12 remained unchanged within the residence time
of 0.03 s. With a longer residence time (0.30 s), a significant
amount of 12 was consumed to give the dialkylation product
13. As discussed in the previous section, the monoalkylation
product 12 seems to be protonated during the course of its
formation to decrease its reactivity because of the electron-
withdrawing effect of the protonated carbamate group (Scheme
12). These experiments using the microflow system revealed
that the reaction of protonated 12, which should be formed by
the reaction of 11, is much slower than the reaction of 11.
The present observations clearly indicated that the second
alkylation is much slower than the first alkylation, presumably
because of the protonation of the first alkylation product.
Therefore, the selectivity observed for the macroscale batch
reactor contradicted the prediction based on the normal kinetics
and could be the disguised chemical selectivity. The selectivity
observed with an IMM micromixer seems to be consistent with
the kinetics.
mixer
lamination width (
µ
m)
P1 (% yield)
P2 (%yield)
P1:P2
ideal mixing
94.6
60.6
89.6
94.7
2.7
19.7
4.5
97:3
75:25
95:5
97:3
a
b
c
100
25
2.5
2.6
Table 7. Product Yields and Distribution Obtained by CFD
Simulation (k1 ) 105 L/(mol s) and k2 ) 103 L/(mol s))
mixer
lamination width (
µ
m)
P1 (% yield)
P2 (%yield)
P1:P2
ideal mixing
94.6
31.1
67.8
94.5
2.7
34.4
15.9
2.7
97:3
a
b
c
100
25
2.5
47:53
55:45
97:3
Table 8. Product Yields and Distribution Obtained by CFD
Simulation (k1 ) 106 L/(mol s) and k2 ) 104 L/(mol s)
mixer
lamination width (
µ
m)
P1 (% yield)
P2 (%yield)
P1:P2
ideal mixing
94.6
14.5
36.1
90.5
2.7
42.7
31.9
4.7
97:3
a
b
c
100
25
2.5
25:75
53:47
95:5
the molecular diffusion coefficient is 10-9 m2/s, and the viscosity
is 0.00119 Pa s; the initial concentrations of the two reactants
are 0.01 mol/L.
Before investigation of the lamination width effect, the effect
of the relative rate of two reactions (k1/k2) was examined for
ideal mixing. In the case of k1/k2 ) 100, P1:P2 was calculated
to be 97:3, whereas in the case of k1/k2 ) 10, P1:P2 ) 87:13.
The former value is closer to the experimental observation for
the present Friedel-Crafts alkylation. Therefore, hereafter, the
simulations were carried out for the case of k1/k2 ) 100. The
results are summarized in Tables 6-8.
CFD Simulation of Competitive Consecutive Reaction
using Micromixing. To obtain a deeper understanding of the
present dramatic effect of micromixing on the selectivity of the
Friedel-Crafts alkylation, it is necessary to analyze how mixing
can influence the selectivity of chemical reactions. Thus, we
carried out CFD (computational fluid dynamics) simulation
using FLUENT. The calculation was performed using a laminar
flow and finite-rate model in the package.20
As shown in Table 6, when k1 ) 104 L/(mol s), the use of a
micromixer of 100 µm lamination width results in the formation
of P1 in ca. 60% yield and P2 in ca. 20% yield. A decrease of
the lamination width causes the increase of the yield of P1 with
the expense of the yield of P2. These results indicate that the
selectivity increases with the decrease of the lamination width.
The increase of the absolute values of k1 and k2 causes
dramatic changes of the selectivity of the reaction in mixer a.
When k1 ) 105 L/(mol s), P1 and P2 formed in almost equal
amounts (Table 7). When k1 ) 106 L/(mol s), the yield of P2
was much larger than that of P1 (P1:P2 ) 25:75), even if k1 is
much larger than k2 (k1/k2 ) 100) (Table 8). Presumably, the
reaction time and mixing time are comparable, and the selectiv-
ity seems to be disguised. The decrease of the lamination width
gave rise to a remarkable increase of the selectivity favoring
Let us consider the competitive consecutive isothermal
reaction, described in Scheme 1 with k1 ) 104, 105, and 106
L/(mol s), whereby 0.01 mol/L solutions of A and B are fed to
a micromixer at a constant rate. As to the shape and size of a
micromixer, three models shown in Figure 6 were used for
simulation. In mixers a-c, A and B are introduced as laminar
flow, the widths (w) of which are 100, 25, and 2.5 µm,
respectively. The flow rate is 0.1 m/s, and the length of the
reactor is 1 m. We also simulated ideal mixing where A and B
are mixed completely immediately after they are introduced to
the mixer. The assumptions of the reactant fluids used in CFD
simulations are as follows: physical properties of two reactant
fluids are the same, that is, the density is 1.317 × 103 kg/m3,
(20) Aoki, N.; Hasebe, S.; Mae, K. Chem. Eng. J. 2004, 101, 323.
9
J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005 11673

A R T I C L E S
Nagaki et al.
120 min at 60 °C, and iodo(trimethylsilyl)methane (42.6 g, 199 mmol)
was added at 0 °C. The reaction mixture was stirred for 14.5 h at 60
°C and poured into water (300 mL). The organic phase was separated,
and the aqueous phase was extracted with ether (300 mL × 3). The
combined organic phase was dried over Na2SO4, and the solvent was
removed to give the crude product. After purification with flash
chromatography (hexane/ethyl acetate 10:1), the product was further
purified by distillation (120 °C, 20 mmHg) to obtain the title compound
t
(36.0 g, 76%): GC R 6.92 min (OV-17; 0.25 mm × 25 m; oven
temperature, 100 °C; rate of temperature increase, 10 °C/min); TLC Rf
0.54 (hexane/ethyl acetate 5:1); H NMR (300 MHz, CDCl3) δ 0.03
Figure 7. Concept of flash chemistry.
1
(s, 9H), 0.90 (t, 3H, J ) 7.2 Hz), 1.16-1.36 (m, 2H), 1.39-1.55 (m,
2H), 2.68 (s) and 2.72 (s) (total 2H, two rotamers), 3.06-3.26 (m,
2H), 3.64 (s, 3H); 13C NMR (75 MHz, CDCl3) δ -1.8 and -1.6, 13.7,
19.8, 29.4 and 29.8, 37.9 and 38.6, 48.7 and 49.3, 51.9 and 52.3, 156.6;
IR (neat): 2957, 1701, 1473, 1225 cm-1; LRMS (EI) m/z 217 (M+),
202 (M+ - CH3), 186 (M+ - OMe), 158 (M+ - CO2Me); HRMS
(EI) calcd for C10H23NO2Si (M+): 217.1498, found: 217.1501.
Generation of Iminium Cation Pool. Typical Procedure. The
anodic oxidation was carried out in an H-type divided cell (4G glass
filter) equipped with a carbon felt anode (Nippon Carbon JF-20-P7,
ca. 320 mg, dried at 250 °C/L mmHg for 1 h before use) and a platinum
plate cathode (60 mm × 30 mm). In the anodic chamber was placed a
solution of 1 (610.3 mg, 2.81 mmol) in 0.3 M Bu4NBF4/CH2Cl2 (56.0
mL). In the cathodic chamber were placed 0.3 M Bu4NBF4/CH2Cl2
(56.0 mL) and trifluoromethanesulfonic acid (1.05 g, 7.00 mmol). The
constant current electrolysis (30 mA) was carried out at -78 °C with
magnetic stirring until 2.5 F/mol of electricity was consumed.
General Procedure of the Reaction in a Macroscale Batch
Reactor. To a solution of an aromatic compound (0.417 mmol) in CH2-
Cl2 (10 mL, cooled at -78 °C) was added a cation pool 2 (10 mL,
cooled at -78 °C) generated from 1 (0.500 mmol) by a syringe pump
(flow rate, 5.0 mL/min). Immediately after mixing, the reaction was
quenched by Et3N (1 mL) at the same temperature. The solvent was
removed under reduced pressure, and the residue was quickly filtered
through a short column (10 cm) of silica gel to remove Bu4NBF4. The
silica gel was washed with ether (300 mL). The combined solution
was concentrated to give a crude product, which was purified by flash
chromatography.
the formation of P1. Even if k1 ) 106 L/(mol s), P1 is formed
in high selectivity (P1:P2 ) 95:5) in mixer c of 2.5 µm
lamination width. The effect of the increase of k1 and k2 obtained
by the simulation is consistent with the experimental observation
that the effect of micromixing is significant only for the reaction
with highly reactive aromatic compounds, for which k1 and k2
are expected to be very large (Table 4).
The simulation data described in Tables 6-8 clearly show
the dramatic effect of the lamination width on the selectivity
of the reaction. Although the micromixer having a 25 µm
channel width was used for the experiments, the simulation
suggests that the lamination width of 2.5 µm is necessary to
realize the observed selectivity. In the IMM micromixer, the
outlet channel is narrowed immediately after the multilamination
configuration has been established. This may cause a decrease
of the effective lamination width. Anyway, the present results
of simulation are, at least qualitatively, consistent with the
experimental observations and clearly indicate the importance
of micromixing for the improvement of the selectivity of
competitive consecutive reactions.
Conclusions
The present study nicely illustrates that the phenomenon of
disguised chemical selectivity is observed for extremely fast
competitive consecutive reactions such as Friedel-Crafts reac-
tions of highly reactive aromatic compounds with N-acylimin-
ium ion pools. The present study also speaks well for the
potentiality of micromixing to effect extremely fast chemical
transformations that are difficult to perform in a controlled
manner using macroscale batch reactors. Micromixing can
provide a solution to the problem of disguised chemical
selectivity that complicates the reactions using more conven-
tional chemical methods.
In conventional organic synthesis, rather slow reactions have
been used because fast reactions are difficult to control. Reaction
times usually range from minutes to hours. To achieve synthesis
in a highly time efficient manner, however, the use of much
faster reactions in a highly controlled way is strongly needed
(reaction time ranges from milliseconds to seconds). Microsys-
tems are expected to serve as powerful tools for conducting
extremely fast reactions in a highly controlled manner to effect
flash chemistry, where a substrate and a highly reactive reagent
are allowed to react to give a desired product quickly (Figure
7).
General Procedure of Micromixing. A solution of nucleophile
(0.417 mmol) in CH2Cl2 (10 mL, cooled at -78 °C) and a cation pool
2 (10 mL, cooled at -78 °C) generated from 1 (0.500 mmol) were
simultaneously introduced to the micromixer, which was dipped in a
coolant at -78 °C, using syringe pumps (flow rate is 5.0 mL/ min).
Then, the reaction mixture coming out from the outlet of the micromixer
was collected with a round-bottomed flask and immediately quenched
by Et3N (1 mL) at -78 °C. The solvent was removed under reduced
pressure, and the residue was quickly filtered through a short column
(10 cm) of silica gel to remove Bu4NBF4. The silica gel was washed
with ether (300 mL). The combined solution was concentrated to give
a crude product, which was purified by flash chromatography.
1-(N-Butyl-N-methoxycarbonylaminomethyl)-2,4,6-trimethoxy-
benzene (12) and 1,3-Bis(N-butyl-N-methoxy-carbonylaminom-
ethyl)-2,4,6-trimethoxybenzene (13). The title compounds were
synthesized by the reaction of a cation pool 2 (10 mL) generated from
1 (0.500 mmol) with 1,3,5-trimethoxybenzene 11 (70.1 mg, 0.416
mmol) by micromixing (IMM single mixer, -78 °C). Yields were
t
determined by GC. 12: 92% yield (GC R 11.7 min, OV-1; 0.80 mm
× 1 m; oven temperature, 100 °C; rate of temperature increase, 10
°C/min). TLC Rf 0.49 (hexane/ethyl acetate 1:1); 1H NMR (300 MHz,
CDCl3) δ 0.84 (t, 3H, J ) 7.2 Hz), 1.11-1.30 (m, 2H), 1.35-1.48 (m,
2H), 2.93-3.08 (m, 2H), 3.63-3.80 (m, 9H), 3.81 (s, 3H), 4.45-4.58
(m, 2H), 6.10 (s, 2H); 13C NMR (75 MHz, CDCl3) δ 13.8, 19.9, 29.5
and 30.1, 37.7 and 37.8, 44.2 and 44.7, 52.1, 55.2, 55.5, 90.1, 105.8,
Experimental Procedures
N-Methoxycarbonyl-N-(trimethylsilylmethyl)butylamine (1). To
a suspension of sodium hydride (60 wt % dispersion in mineral oil,
7.91 g, 197 mmol) in DMF (200 mL), N-(methoxycarbonyl)butylamine
(21.6 g, 165 mmol) was added at 0 °C. The mixture was stirred for
156.8, 156.0, 160.8; IR (neat) 2957, 1697, 1608, 1498, 1151 cm-1
LRMS (EI) m/z 311 (M+), 280 (M+ - OMe), 252 (M+ - CO2Me),
;
9
11674 J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005

Competitive Consecutive Reactions using Micromixing
A R T I C L E S
181 (M+ - NBuCO2Me); HRMS (EI) calcd for C16H25NO5 (M+):
311.1733, found: 311.1733. 13: 4% yield (GC R 16.7 min, column,
the final stage, Et3N (8 mL/min) was introduced through M2 to quench
the reaction. For M2, Yamatake YM-1 was used.
t
OV-1; 0.80 mm × 1 m; oven temperature, 100 °C; rate of temperature
increase, 10 °C/min). TLC Rf 0.37 (hexane/ethyl acetate 1:1): 1H NMR
(300 MHz, CDCl3) δ 0.81 (t, 6H, J ) 7.2 Hz), 1.09-1.28 (m, 4H),
1.29-1.43 (m, 4H), 2.86-3.07 (m, 4H), 3.59-3.79 (m, 9H), 3.80 (s,
6H), 4.50 (s) and 4.55 (s) (total 4H, two rotamers), 6.23 (s, 1H); 13C
NMR (75 MHz, CDCl3) δ 14.0, 20.2, 29.8 and 30.3, 38.8, 44.5 and
45.1, 52.6, 55.8, 62.6, 91.3, 111.2, 157.1, 160.0; IR (neat) 2957, 1698,
1609, 1152 cm-1; LRMS (EI) m/z 454 (M+), 423 (M+ - OMe), 395
(M+ - CO2Me); HRMS (EI) calcd for C23H38N2O7 (M+): 454.2679,
found: 454.2682.
Estimation of Relative Rates of the First Alkylation and Second
Alkylation using the Microsystem. A microsystem consisting of two
micromixers (M1 and M2) and a tube reactor (R1) was used. The whole
system was cooled in a dry ice bath at -78 °C. For M1, where the
cation pool and an aromatic compound were mixed, a splitting and
recombination type micromixer, Yamatake YM-1 was used. Thus, a
solution of 2 (0.05 M) generated from 1 and that of an aromatic
compound (0.03 M) were introduced to M1 by the syringe pumping
technique (flow rate: 8.0 mL/min) at -78 °C. Then, the reaction
mixture was introduced to a tube reactor (R1) (φ ) 1.0 mm, X cm). In
DFT Calculations. The DFT calculations were carried out at the
B3LYP/6-31G(d) level using the Gaussian 2003W, Revision-B.05.21
Geometries were fully optimized. All the optimized structures were
local minima according to the vibration analysis. The transition state
was verified to have exactly one negative eigenvalue. A negative
eigenvalue was animated using Gaussian View 3.0 and proved to be
the transition state for formation of the expected bond.
Acknowledgment. This work was partially supported by a
Grant-in-Aid for Scientific Research and the Project of Micro
Chemical Technology for Production, Analysis, and Measure-
ment Systems of NEDO, Japan.
Supporting Information Available: Experimental procedures,
spectroscopic data of compounds, and geometries obtained by
DFT calculations (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
JA0527424
(21) Pople, J. A. et al. Gaussian 03, Revision B.05; Gaussian, Inc.: Pittsburgh,
PA, 2003.
9
J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005 11675